A Technical History of the SEI
ihQTwP
ihQTwP
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
Measurement and Analysis<br />
The Challenge: Measuring S<strong>of</strong>tware Development Capabilities and<br />
Products<br />
Measurement and analysis in s<strong>of</strong>tware engineering has long been a topic <strong>of</strong> interest. Without it,<br />
<strong>the</strong>re is no clear, quantitative picture <strong>of</strong> s<strong>of</strong>tware development capabilities or a basis for predicting<br />
or comparing products and processes. Although early s<strong>of</strong>tware projects used several systems for<br />
predicting s<strong>of</strong>tware costs, measurement was done in different ways and was <strong>of</strong>ten based on definitions<br />
that were inconsistent. The need for a standard and reliable set <strong>of</strong> measures that would help<br />
acquisition program managers and s<strong>of</strong>tware development contractors alike was an early priority<br />
for <strong>the</strong> <strong>SEI</strong>.<br />
A Solution: Approaches for Collecting and Analyzing Data<br />
In response to <strong>the</strong> DoD’s 1991 S<strong>of</strong>tware Technology Strategy, <strong>the</strong> <strong>SEI</strong> agreed to lead <strong>the</strong> development<br />
<strong>of</strong> a set <strong>of</strong> core measures to “help <strong>the</strong> DoD plan, monitor, and manage its internal and contracted<br />
s<strong>of</strong>tware development projects” [Carleton 1992]. In collaboration with measurement experts,<br />
including those who developed prediction systems, <strong>the</strong> <strong>SEI</strong> developed definition<br />
frameworks for a set <strong>of</strong> core measures. The measures focus on size, defects, effort, and schedule.<br />
These definition frameworks make it possible for organizations to use <strong>the</strong> measures that best<br />
match <strong>the</strong>ir processes and infrastructure while benefiting from a standard way <strong>of</strong> describing <strong>the</strong><br />
operational definitions in detail.<br />
Once <strong>the</strong> definition problem for measures was resolved, <strong>the</strong> <strong>SEI</strong> turned its attention to helping organizations<br />
decide what to measure. Experience from management information systems shows<br />
that many reports could be generated that have little to no impact on decision making within <strong>the</strong><br />
organization. The <strong>SEI</strong> began to investigate <strong>the</strong> goal-question-metric method for aligning measurement<br />
with information needs in <strong>the</strong> organization [Basili 1984]. The <strong>SEI</strong> modified this approach to<br />
include explicit consideration <strong>of</strong> <strong>the</strong> output <strong>of</strong> <strong>the</strong> measurement activity—that is, <strong>the</strong> indicator to<br />
be used by decision makers. The method was dubbed goal-question-indicator-measure (GQ(I)M).<br />
Using <strong>the</strong> GQ(I)M approach helps mitigate <strong>the</strong> risk <strong>of</strong> measuring and reporting information that<br />
provides little or no value to <strong>the</strong> organization.<br />
As approaches for conducting measurement effectively matured and were disseminated, <strong>the</strong> next<br />
significant challenge became data analysis [Paulk 2000]. The notion <strong>of</strong> analysis, especially analysis<br />
related to process improvement, was <strong>of</strong>ten equated with <strong>the</strong> high-maturity practices <strong>of</strong> <strong>the</strong> SW<br />
CMM and CMMI. More focus on <strong>the</strong> “analysis” part <strong>of</strong> measurement and analysis was also<br />
spurred by <strong>the</strong> establishment <strong>of</strong> <strong>the</strong> Measurement and Analysis process area in <strong>the</strong> CMMI [CPT<br />
2002].<br />
An early and foundational work in this area was Measuring <strong>the</strong> S<strong>of</strong>tware Process, which showed<br />
how statistical process control techniques could be fruitfully applied to s<strong>of</strong>tware data [Florac<br />
1999]. This was followed by substantial work on <strong>the</strong> application <strong>of</strong> Six Sigma analytical techniques<br />
to s<strong>of</strong>tware engineering [Penn 2007]. The Six Sigma connection provided a rich set <strong>of</strong><br />
tools as well as a “brand” that already had roots in many organizations, facilitating its adoption.<br />
The current work on developing estimates early in <strong>the</strong> DoD acquisition lifecycle incorporates<br />
CMU/<strong>SEI</strong>-2016-SR-027 | SOFTWARE ENGINEERING INSTITUTE | CARNEGIE MELLON UNIVERSITY 149<br />
Distribution Statement A: Approved for Public Release; Distribution is Unlimited.